Abundance is no guarantee of a resilient society
Two centuries of extraordinary growth have taught us to expect more. Aaron Bastani asks whether that expectation has left us unprepared for disruption.
The story of humanity over the last 250 years, shaped by the steam engine, a faith in markets and the prodigious consumption of fossil fuels, is one of increasing abundance.
Take the planet’s population. In 1800 it stood on the cusp of one billion. By 2022, humanity had surpassed eight. There’s health and longevity too: in the 2020s the average person can expect to make it to their 70s, while infant mortality has collapsed. And, perhaps most visibly, it pertains to a vast profusion of stuff—an “immense accumulation of commodities,” as one German philosopher memorably put it.
It’s worth underscoring how spectacular this change has been. While it took 300,000 years for Homo sapiens to reach its first billion, that figure doubled again in a little more than a century. Thereafter things only accelerated, with the population doubling once more in less than 50 years. Such explosive change had never happened before and is unlikely to be repeated.
Environmental scientists have termed this development—-of a booming population alongside unprecedented consumption—the “Great Acceleration.” Beginning around the middle of the last century, a dramatic surge in human activity had profound consequences for almost all Earth systems, from energy and water consumption to marine fish capture. For each of these areas, and more than a dozen more, the graphs after 1945 resemble a hockey stick.
It’s unsurprising, then, that we take rising abundance for granted. It’s the only reality almost anyone alive has ever known. That’s also why warnings about systemic risk—whether from climate change, faltering democracy or AI—can struggle to cut through. For all the claims that ours is an era of rising political chaos, we also suffer from a recency bias. For eight decades, we’ve mostly known only growth. We presume, therefore, that the future will look much the same.
Conversations at the frontier, in your inbox
SubscribeAlongside that temporal myopia sits a deeper truth: humans are built for scarcity. It’s why we prefer that buttery pastry to a beetroot salad—for hundreds of millennia, all we knew was deprivation. The drive for more, then, shouldn’t necessarily be seen as greed, but as rational self-interest wired for a different time. Humans are meant to have insatiable needs, precisely because in the past, we had no way of meeting them.
That same tension, between abundance and scarcity, runs through modern economics too—a discipline that emerged alongside the Steam Age. Adam Smith’s “The Wealth of Nations,” which popularized the notion that self-interest can also serve the public good, was published in 1776, the same year as the Declaration of Independence. Remarkably, 1776 was also when Watt and Boulton’s first steam engines entered service. It’s a useful date, then—precisely 250 years ago—to pin as the start of the modern world.
The most vocal champions of AI see their field in two ways: as an extension of that 250-year story, and as a transformation on par with the Industrial Revolution. If the 1770s marked the take-off moment for the “First Machine Age,” to borrow a phrase from Erik Brynjolfsson and Andrew McAfee, then we are now in the foothills of a second. It’s in this context that, for some, the central question is how far this second age might go—and whether the real takeoff begins once a machine is capable of augmenting its own intelligence. Achieve AGI, more radical voices intone, and humanity, in some spheres of economic life, edges toward post-scarcity. This poses a challenge to economic theory, and how its students make sense of the world. After all, the discipline’s focus is to study the allocation of scarce resources among competing, unlimited ends—that impulse we inherited from a world of deprivation and cave paintings.
AI and the polycrisis
Rather than raising questions about market systems in a world of AGI, or how automation might compress wages, I want to probe a less examined tension between abundance, scarcity and resilience. Because while ours is an age of (uneven) plenty, it is also infused by a sense of crisis. That’s partly a function of the media environment, from 24-hour news to smartphones, but it goes deeper too. For one thing, planetary conditions are changing, with global average surface temperatures increasing by about 0.2 degrees Celsius a decade, and ExxonMobil recently projecting a rise as high as 3 degrees Celsius above preindustrial levels by the middle of this century. That’s happening alongside rising geopolitical tensions and the demise of the relatively benign order that followed the Cold War.
Then there’s the increasingly polarized politics of the West, as electorates punish elites and reward populist insurgents. That’s before touching on the challenges of AI, existential or otherwise, which may prove the greatest test of them all. Indeed, a plausible “middling” outcome, far removed from the doomsday scenarios favored by many, is that LLMs do to white-collar work what Netflix did to Blockbuster.
The claim that we live in a moment of crisis is hardly new. What is novel, however, is the widespread agreement behind such a claim. In early 2026, nearly two-thirds of the British public expressed disappointment in their government, according to the pollster Ipsos. the country as heading in the wrong direction according to the pollster Ipsos. Meanwhile, according to a recent poll from the Emerson College, 41% of Americans are pessimistic about the future, compared to just 15% fifty years ago.
Skepticism about the future is spreading, alongside a loss of trust in authorities’ ability to fix things. Some, like the historian Adam Tooze, see this as emblematic of a “polycrisis,” in which disparate global challenges don’t just coexist, but feed one another. Here, geopolitical breakdown drives energy and price inflation. That inflation fuels a cost-of-living crisis that never quite ends. And that, in turn, feeds domestic polarization—which makes decarbonization, or reforming areas like elderly care, impossible.
That is just one chain reaction among many. Another is automation, particularly the rise of AI as a frontier technology. This feeds the polycrisis by deepening the sense of confrontation between the U.S. and China. At the same time, LLMs threaten to further compress most workers’ wages, exacerbating inequality and in turn feeding division across the West.
In other words, AI is as much a potential source of vulnerability as it is an opportunity.
In other words, AI is as much a potential source of vulnerability as it is an opportunity. And because of how the polycrisis works, it can’t be understood in isolation from issues like climate change, populism or demographic aging.
The lessons of Icarus
Which brings us to abundance. In the age of polycrisis, a once counterintuitive idea feels increasingly true: wealthy societies, those first pioneers of abundance, aren’t just failing to solve their social and political problems. They’re also struggling to generate new wealth.
For the academic John Rapley, these two issues are connected. Borrowing from business literature (and Greek mythology), he describes the phenomenon as “Icarus economics,” a situation where a nation’s wealth becomes such that it “triggers a cycle of behavioral and social changes whose effect … is to progressively inhibit a society’s economic dynamism.” Over time this means that the accumulation of wealth “creates demands for its preservation that increasingly weigh on (further) growth.” Interest groups who presently benefit pursue sectional interests at odds with the creation of new value. Growth invariably leads to degrowth.
Rapley argues that wealth changes societies just as it changes people, making them slower, less agile and more risk-averse. More than two centuries of growing wealth has actually become a source of weakness, which means those pushing AGI as a route to abundance have a question to answer: would its deployment, steered by the very interests already described, exacerbate or mitigate such dynamics?
But is Rapley’s provocative hypothesis actually true? Regarding resilience, the experience of COVID-19 suggests so. In the pandemic’s opening months, most presumed it would be poorer countries, with the weakest health infrastructure, that would suffer most. Countries like Britain and the U.S., meanwhile, were better placed because of their wealth.
Except that’s not what happened. Instead, Britain and the U.S. experienced some of the worst pandemic-related outcomes of any country—precisely because of their deep integration into a global system that buckled almost immediately. Outsourced production, until recently venerated as a model of efficiency, became a source of weakness. Just-in-time delivery, which had replaced the “just in case” caution of decades past, similarly fell apart.
Then there’s the claim of stagnation. Few would disagree that Rapley’s diagnosis applies to Europe’s economies since 2020. But surely things are different with the U.S.? After all, the conventional wisdom says that while the old world flounders, the new one prospers.
The truth is rather more complicated. Because while American GDP has increased by almost $7 trillion since 2022, the country’s national debt surged by more than $9 trillion over the same period. In other words, U.S. growth is now a function of high deficit spending.
This appears to confirm that our model is struggling to generate new wealth while also being unable to meet challenges. Consider, for instance, Europe’s inability to ramp up shell production to support Ukraine, or America’s failure to deal with Iran.
Advocates for AGI might accept the first part, about sclerotic growth, and say that is more reason to double down on technological disruption. After all, the world before the steam engine was relatively stagnant too. Pursuing greater abundance through technology, as an end in itself, is the best card we have. I disagree. Instead, we should focus on resilience—human, social and technological—with this in turn shaping the kind of AI we build and deploy. Because without such considerations, and by simply pursuing “abundance,” we will achieve anything but—exacerbating the most obvious inputs feeding the polycrisis, inequality chief among them.
AGI will make inequality, and the polycrisis, worse
Is inequality getting worse? Unequivocally, yes.
Is inequality getting worse? Unequivocally, yes.
At the turn of the millennium, Bill Gates was the world’s wealthiest person, boasting a fortune of $60 billion ($110 billion in today’s prices). By June of this year, however, Elon Musk became a trillionaire. In Britain, the combined fortune of the country’s 200 wealthiest families stands at 20% of GDP. As recently as 1994, that figure was just 6%. In France, meanwhile, the country’s 500 most affluent families enjoyed a combined wealth equivalent to 6% of national output in the 1980s. Most recently, that number is 42%.
How this happened is contested. A decades-long collapse in interest rates is part of it, as are weaker trade unions. But whatever your preferred cause, it’s beyond a doubt that spiraling inequality is feeding the polycrisis like little else. So if most of the new wealth created by AI goes to a tiny section of society, as I think it would with AGI, one can reasonably presume that it would inject more chaos into an already destabilizing system.
Economic inequality, most notably on a regional level, generates political disruption. Given that Donald Trump and MAGA were initially powered by small-town America, it’s worth reflecting on how the percentage of personal income going to small, nonmetropolitan U.S. towns has, according to the economist Daron Acemoglu, halved since 1969 (from 21% to 10.5%). How low is that figure allowed to go? And would AGI accelerate or reverse the demise of such places? Simply put, the main concern regarding AI-led abundance should be whether it would reduce or aggravate inequality. If it’s the latter, we risk a societal death spiral, with the populist uprisings of the last decade insignificant compared with what’s in store.
Right now the evidence indicates that widespread adoption of AI (particularly, with the emergence of AGI) would serve to deepen inequality between households, regions and even nations. That’s primarily the result of AI substituting for labor, with the returns flowing to the owners of capital rather than to those who work. Then there’s the “winner-takes-all” dynamic of the existing AI industry, where a small number of trillion-plus-dollar companies mostly based in the U.S. control the frontier models, compute infrastructure and data. This would likely concentrate wealth even more than previous technological revolutions did, since AI has unusually strong network effects and economies of scale.
Finally, there’s geopolitics. Compute, chips and frontier AI development overwhelmingly sit in the hands of two “AI superpowers”: the U.S. and China. That raises the specter of a historic divergence, with AI’s leaders pulling away from its laggards, much as the West did at the start of the 19th century when its early mastery of steam technology became the foundation for “gunboat diplomacy.” It’s a question worth putting to those beyond Washington and Beijing: Do we want this kind of gap, along with the forms of political and economic dependence that would ensue? What would a stable state system even look like under such conditions?
Pursuing resilient abundance
In the first issue of The Humanist Review of AI, Daron Acemoglu made the case for the AI industry to move away from “the idea of superhuman machines,” adding that we need to “recognize that a better future lies in cultivating human skills, not sidelining them.” Thus, rather than seeking to create AGI and eliminate jobs, we should pursue technology that expands the productivity of people in work. In any debate focusing on abundance and resilience, the significance of this is obvious. If widening inequality is a leading cause of the polycrisis, then AI should only be promoted if it narrows the divide, not deepens it. The call for pro-worker AI is not, therefore, merely an ethical preference—where we strive to create better, more meaningful jobs—but is also the first step, among several, in pursuing resilient abundance.
Alongside the enhanced social resilience of pro-worker AI and a focus on deploying technology to reduce inequality—which also requires alternative forms of ownership—is the need to pursue human resilience and rethink our personal relationships to technology.
For all their benefits, there’s growing evidence that social media and the mobile internet have also brought greater isolation, loneliness and declining mental health. Besides such grim trends, however, are the political implications, with the digital media environment fueling polarization and popular anger. As the columnist and writer James Marriott put it in his recently published “The New Dark Ages”, AI is liable to make this even worse with generative AI, in particular, representing a step toward a post-literate society imbued with “mystical, tribal and autocratic” characteristics, reminiscent of our “pre-literate past.” While the era of TV entertainment was often blamed for a culture of apathy, Marriott’s concern is that our public sphere is becoming outright chaotic and illogical. Post-literate publics, he argues, are losing the capacity for rational deliberation—the very thing, however flawed, that made mass democracy possible.
For all the inequities of the First Machine Age, which began about a quarter of a millennium ago, it was attended by growing literacy and, as a result, an increasingly formidable public sphere. That, in turn, powered new forms of association, from reading clubs and unions to political parties, which consequently reprogrammed the Industrial Age from a world of stagnant living standards (termed the “Engels pause”) to one of public works, falling infant mortality and rising wages. With that in mind, pursuing economic “abundance” through AGI, at the expense of producing literate, self-reflective citizens, seems to be a very bad idea.
Instead, as we stand on the brink of a remarkable technological dawn, we should deploy AI while building human resilience. That requires reasserting the primacy of humanist culture—something at odds with the political orthodoxies of recent generations. In his seminal work “Technopoly”, the media theorist Neil Postman contrasts the modern era with the Middle Ages, claiming the tools of our ancestors “did not attack (or more precisely were not intended to attack) the dignity and integrity of the culture into which they were introduced.” The opposite has applied for over a century, where technological innovations—from the car and television to the smartphone and now AI—were allowed to re-configure society. Postman puts this better still when he writes about how the idea of human progress has been replaced by that of technological progress, the aim no longer being to “reduce ignorance, superstition and suffering but to accommodate ourselves to the requirements of new technologies.” This is our world—of employee training instead of education and politicians and entrepreneurs who claim that trying to “stop progress” is futile. It’s the fatalistic, semi-religious mindset that dominates thinking around AGI.
The first step towards a meaningful response is to reaffirm what constitutes the Good Society. Which in turn requires a shift in our relationship to digital tools, such as banning smartphones in schools, no social media accounts before 16, potential limits on screen time, eradicating short-form video. All are good policies. More broadly, though, we need to grasp the dangers of shifting toward a post-literate social order and act accordingly. That means returning to an emphasis on the liberal arts and restating the relationship between education, culture and freedom. In recent decades, the humanities have been denigrated, their value dismissed when compared to STEM subjects.
In a world of increasingly powerful, productive computation, we will, paradoxically, need to be more attuned to human flourishing.
But the point of the liberal arts is to educate us in how to achieve freedom: to think rationally for ourselves, to control our appetites and to participate in collective self-government. To regard such training as the preserve of elites is, essentially, to give up on mass democracy altogether. Defending the literate society, and the humanities, should therefore be a core ambition as we grapple with a world re-made by thinking machines. This is not a nice to have, but a cornerstone of crafting a resilient society body politic. In a world of increasingly powerful, productive computation, we will, paradoxically, need to be more attuned to human flourishing.
Finally comes technological resilience. Because without a national tech stack, the sovereignty of countries will become meaningless. As Huawei founder Ren Zhengfei told Jiang Zemin, then the CCP’s General Secretary, in 1994, “a country without its own program-controlled switches is like one without an army.” Over time that pithy formulation will apply to at least some semiconductor manufacturing, as well as data centers, cloud computing and, most critically, AI models.
Of course national industry and a strong sense of agency have always gone together. But the importance of AI as a technology is an order of magnitude greater than an industry like steel, which, however important, could be emulated in a decade or two by Stalin’s Russia or Meiji-era Japan. AI, by contrast, is a general-purpose technology closer to steam or electricity. Again, this matters from the perspective of resilience and the polycrisis, because whatever its various faults over the last 250 years, the nation-state has been the leading repository of democratic accountability and global problem-solving. Depriving most countries of technological sovereignty, one suspects, would lock in ever wider state failure. I would go as far as saying that the perpetuation of the state system, which has endured for almost four centuries—and was the cultural software allowing the First Machine Age to flourish—is at stake.
Complexity, collapse and resilience
High technology. An elite culture that transcends borders. Global interconnectedness and rising trade. You might not know it, but we’ve been here before.
Around 1400 B.C., Akkadian was a diplomatic language used across Egypt, parts of Greece, Cyprus, Anatolia and as far as Iran. This bound together a geopolitical order—a brotherhood of kings—with Egyptian, Hittite, Elamite and Mycenaean rulers recognizing one another’s legitimacy. This was, then, an early form of globalization, with logistical routes and a shared cultural foundation allowing unprecedented affluence.
And yet this order, built on diplomacy and sophisticated technologies for the time, such as chariot-building and metallurgy, unraveled. Whole societies, like the Hittite Empire, crumbled while cities such as Ugarit, the Singapore of its day, were destroyed and never resettled. For the Mycenaean Greeks, the disruption was so great that literacy itself was abandoned for four centuries. Today we call that terrifying episode the Bronze Age Collapse.
We can’t be entirely sure what destroyed those societies, but it was likely a combination of climate change, natural disasters and invasions. Underpinning it all, though, was a system whose complexity had suddenly become a source of weakness. Societies whose wealth required stability and trade were exposed as that order evaporated.
The Bronze Age Collapse, therefore, was a polycrisis. And, as in our own time, it was the most advanced societies, at the forefront of what could be achieved by humans, that were the most vulnerable. Mycenaean Greece, home to monumentalist architecture that would be unsurpassed for centuries, foundered to such an extent that later generations believed only the mythical Cyclopes could have built the gigantic walls of Mycenae and Tiryns. Linear B, the archaic script they used, disappeared. This was Icarus Economics 1.0.
Those are the stakes we face in our own time. Get it wrong with AI, especially AGI, and we may contribute to a series of feedback loops—including runaway climate change and a collapse in the geopolitical order—that leave us further back than where we started. What was true over three millennia ago also applies today: wealth without stability, and technological sophistication without political equilibrium, is a recipe for disaster. Pursue abundance by all means, but not at the cost of the very civilization, and social contract, underpinning it. And for that we need to accept the primacy of resilience: not as a brake on abundance, but as its safeguard across the generations.